The Abortion–Breast Cancer Connection potx

303
The Abortion–Breast Cancer
Connection
Joel Brind
The reputation of abortion as safe for women—which claim is explicitly part
of the Roe v. Wade decision—has rightfully come under serious question for many
reasons over the years since Roe. One of the reasons that “safe abortion” has come
under question is the evidence linking abortion to an increased risk of breast cancer
(ABC link). The ABC link has been an issue that has been in and out of the limelight
in recent years. It is an issue which has stubbornly refused to go away despite
recurrent pronouncements from high places of its nonexistence.
A recent example is a 2004 article in the prestigious British medical journal
The Lancet.
1
The paper was promoted by the mainstream media as “a full analysis
of the current data.”
2
According to the byline on the paper, the results of all these
studies were compiled into a “collaborative reanalysis,” by the “Collaborative Group
on Hormonal Factors in Breast Cancer,” a group of authors too numerous to list.
However, a small print footnote reveals that the study was actually put together by
a group of five scientists at Oxford University, headed by prominent British epidemi-
ologist Valerie Beral. The Beral study’s conclusion is unequivocal: “Pregnancies
that end as a spontaneous or induced abortion do not increase a woman’s risk of
1
Collaborative Group on Hormonal Factors in Breast Cancer, “Breast Cancer and Abor-
tion: Collaborative Reanalysis of Data from 53 Epidemiological Studies, Including 83,000
Women with Breast Cancer from 16 Countries,” The Lancet 363.9414 (March 27, 2004): 1007–
1016.
2
David Wahlberg, “Study: Breast Cancer Not Tied to Abortion; Group Backs Up

Institute’s Earlier Findings,” Atlanta Journal-Constitution (GA), March 26, 2004, A9.
THE NATIONAL CATHOLIC BIOETHICS QUARTERLY \ SUMMER 2005
304
developing breast cancer.”
3
This conclusion is remarkably reminiscent of the Na-
tional Cancer Institute’s (NCI) statement given on its “Cancer Facts” web page on
“Abortion, Miscarriage, and Breast Cancer Risk,” carried on the NCI website since
the spring of 2003.
4
On this “fact sheet,” the NCI concludes that “having an abor-
tion or miscarriage does not increase a woman’s subsequent risk of developing
breast cancer.”
The trouble is, to accept this conclusion, one needs to dismiss almost half a
century’s worth of data which do show a significant link between abortion and an
increased risk of breast cancer. Beral et al. suggest that those previous studies
“yielded misleading results,”
5
and that one should trust the largest, most recent
studies (i.e., those which show no ABC link). Such apparently knowledgeable pro-
nouncements seem just a bit too self-assured in an age when concerns about women’s
health reign supreme.
If one can be certain of anything about the ABC link, it is surely that the
question of its very existence is important enough for a careful evaluation, given the
millions who choose abortion and the tens of thousands who die of breast cancer
each year. The present work will therefore examine the ABC link in some historical
and scientific detail, offering a perspective on an issue that is at the center of a long-
running public policy debate that, having been sucked into the maelstrom of the
“abortion wars,” plays out in legislatures and courtrooms and newspaper editorials,
as well as in scientific and medical journals.

Early History of the ABC Link
Neither the ABC link nor the efforts to suppress it are new; the first published
study to document it occurred almost half a century ago. Over the years, denial of
the ABC link has become the party line of all major governmental agencies (includ-
ing the World Health Organization
6
[WHO]), mainstream medical associations (in-
cluding the American College of Obstetricians and Gynecologists
7
and Royal Col-
lege of Obstetricians and Gynaecologists
8
), and the most prestigious medical jour-
nals (including the New England Journal of Medicine
9
).
3
Collaborative Group, “Breast Cancer and Abortion,” 1007.
4
National Cancer Institute, “Abortion, Miscarriage, and Breast Cancer Risk,” http://
cis.nci.nih.gov/fact/3_75.htm.
5
Collaborative Group, “Breast Cancer and Abortion,” 1007.
6
World Health Organization, “Induced Abortion Does Not Increase Breast Cancer Risk,”
Fact Sheet No. 240 (June 2000), />7
American College of Obstetricians and Gynecologists, “ACOG Finds No Link Be-
tween Abortion and Breast Cancer Risk,” news release, July 31, 2003, />from_home/publications/press_releases/nr07-31-03-2.cfm.
8
Royal College of Obstetricians and Gynaecologists, “RCOG Statement on Abortion

and Breast Cancer,” January 28, 2004, />ID=33&PressReleaseID=82.
9
Patricia Hartge, editorial, “Abortion, Breast Cancer, and Epidemiology,” New En-
gland Journal of Medicine 336.2 (January 9, 1997): 127–128.
BRIND \ THE ABORTION–BREAST CANCER CONNECTION
305
A 1957 nationwide study in Japan (published in the English language Japa-
nese Journal of Cancer Research) reported that women who had breast cancer
had a three-fold higher frequency of pregnancies that had ended in induced abor-
tion.
10
As abortion was neither legal nor common in many places, however, such
studies were few and far between. But in 1970, a very high profile, multinational
WHO study, based at Harvard, and published in the WHO’s own Bulletin, reported
a disturbing trend “in the direction which suggested increased risk associated with
abortion—contrary to the reduction in risk associated with full-term births.”
11
The
WHO study findings were not based specifically on induced abortion, including both
induced abortions and miscarriages, but it is interesting that they came out just about
the time when, in the United States and elsewhere, the question of legalization of
induced abortion was being widely considered. The fact that the WHO findings
never entered the debate reveals a disturbing—and continuing—disconnect between
the so-called women’s health advocates pushing for legalized abortion, and any
genuine concern for women’s health.
The first epidemiological study on American women to consider the ABC link
specifically was published in the British Journal of Cancer (BJC) by Malcolm
Pike and colleagues
12
at the University of Southern California in 1981. Since abor-

tion had only been legal in the United States for scarcely a decade by then, the only
appropriate candidates for the study were women diagnosed with breast cancer by
their early thirties. In other words, the subjects needed to have been young enough
to have been exposed to legal abortion. The results of the Pike study made head-
lines: women who had an abortion before they had any children were at a 2.4-fold
(i.e., 140 percent) increased risk for breast cancer.
The response of the scientific community to the Pike study was dichotomous:
reflective of responsible concern from some quarters, and of active denial from
others. Exemplifying the former was a 1982 review in the prestigious journal Sci-
ence by Willard Cates, Jr. of the Centers for Disease Control and Prevention
(CDC).
13
Writing on the overall, roughly decade-long history of the safety of legal
abortion in the United States, Cates expressed his concern: “There is some concern
about … possibly higher risks of breast cancer in certain women.” Exemplifying the
effort to deny the ABC link, however, was a 1982 study published in the BJC by a
group from Oxford University (interestingly, with overlapping authorship with the
2004 Oxford [Beral] “collaborative reanalysis”).
14
The 1982 Oxford study targeted
10
M. Segi et al., “An Epidemiological Study on Cancer in Japan,” GANN (Japanese
Journal of Cancer Research) 48, Suppl. (1957): 1–63.
11
B. MacMahon et al., “Age at First Birth and Breast Cancer Risk,” Bulletin of the
World Health Organization 48.2 (1970): 209–221.
12
M. C. Pike et al., “Oral Contraceptive Use and Early Abortion as Risk Factors for
Breast Cancer in Young Women,” British Journal of Cancer 43.1 (January 1981):72–76.
13

W. Cates, Jr., “Legal Abortion: the Public Health Record,” Science 215.4540 (March
26, 1982): 1586–1590.
14
M. P. Vessey et al., “Oral Contraceptive Use and Abortion before First Term Pregnancy
in Relation to Breast Cancer Risk,” British Journal of Cancer 45.3 (March 1982): 327–331.
THE NATIONAL CATHOLIC BIOETHICS QUARTERLY \ SUMMER 2005
306
Pike’s study specifically, and claimed greater credibility for its much larger number
of patients (1,176 compared with 163 in the Pike study) and much greater age range
(up to age 50, compared with a maximum age of 32 in the Pike study). The Oxford
group’s conclusion was as noteworthy for its emotional tone as for its contrary
result: “The results are entirely reassuring, being, in fact, more compatible with
protective effects than the reverse.” Scientifically, it is a simple matter to explain the
Oxford group’s negative result: It was based almost entirely on miscarriages, as so
few of the women in the study had been young enough to be exposed to legal
induced abortion. The biological differences between these two events are clear,
and will be discussed in some detail a bit later on in the present paper. It was also
particularly telling that, in a paper based entirely on quantitative data, the only quasi-
quantitative expression in the entire text (or tables) for the number of women in the
study who had actually undergone an induced abortion was “only a handful of
women.” Clearly, this Oxford “study” was little more than a fabrication of appar-
ently negative data, designed to “reassure” the public about the safety of abortion.
The Biology behind the Statistics
One would think, especially given the overwhelmingly elective nature of in-
duced abortion, that the precautionary principle would prevail, if not in terms of legal
regulation, then at least in terms of recommendations by medical societies and pub-
lic health agencies. That is to say, even one or two studies showing a significant
association between induced abortion and future breast cancer risk would surely
raise some red flags about the procedure’s safety. Yet not only was a statistical
connection showing up in the vast majority of studies that had examined the issue,

but by the early 1980s, a clear picture of the physiological events explaining that
connection was beginning to emerge.
One important line of evidence providing biological support for the ABC link
came from the field of reproductive endocrinology (the study of the hormones of
reproduction). Only during the 1970s did laboratory methods for measuring such
hormones as estradiol (the main active form of estrogen) and progesterone easily
and cheaply become widely available. In 1976, a landmark study by two Swiss
obstetricians, Kunz and Keller, was published in the British Journal of Obstetrics
and Gynaecology.
15
The Kunz and Keller study documented a clear difference
between the enormous rise of estrogen and progesterone in the first trimester of
viable pregnancies, and the stunted and short-lived rise of these hormones during
pregnancies destined to abort spontaneously (miscarry). These findings dovetailed
perfectly with the patterns of differences in breast cancer risk following different
pregnancy outcomes that was now clearly emerging from the epidemiological data.
During the same period of the late 1970s, key experimental research on labo-
ratory rats was providing another avenue of verification of the ABC Link. Jose and
Irma Russo, a prominent husband-and-wife research team at the Michigan Cancer
15
J. Kunz and P. J. Keller, “HCG, HPL, Oestradiol, Progesterone and AFP in Serum in
Patients with Threatened Abortion,” British Journal of Obstetrics and Gynaecology 83.8
(August 1976): 640–644.
BRIND \ THE ABORTION–BREAST CANCER CONNECTION
307
Foundation in Detroit (they are now at Fox Chase Cancer Center in Philadelphia)
conducted a landmark study in which rats were exposed to standard, cancer-pro-
ducing doses of a known chemical carcinogen after different pregnancy outcomes.
16
Almost 80 percent of rats who had undergone surgical abortion developed breast

cancer (similar to rats not allowed to become pregnant at all), while those allowed a
full-term pregnancy were completely protected from developing the disease.
Not only do experimental animals provide verification of epidemiological data,
but their bodies can be examined microscopically during and after the experiment.
In this way, the Russos have been major players in the discovery of the changes that
take place in the mammalian breast before, during, and after pregnancy. In the
photograph on the following page, lobules type 1 (LOB 1) represent those very
primitive structures present in the breast at birth. Lobules type 2 (LOB 2, not pic-
tured) are present in greatest number after puberty, but before any pregnancy. Only
toward the end of a full-term pregnancy (about 32 weeks gestation in the human
species) do most lobules become lobules type 3 (LOB 3; see accompanying photo-
graph on following page), which are much denser and elaborate, and capable of
lactation. Lobules type 4 are those which are actively producing milk. The progres-
sion from type 1 to type 3 requires an enormous amount of cellular multiplication.
Lobules type 3 are also terminally differentiated, meaning the cells’ ability to multi-
ply has switched off. Part of the development process in the breast, and in most
types of tissue generally, is this terminal differentiation. Cancer, whether of the
breast or any other tissue, is a disease wherein cellular multiplication or proliferation
is out of control. Therefore, it is only those cells which are still capable of prolifera-
tion, such as the cells in lobules type 1 and 2, which are vulnerable to the effects of
carcinogens. Carcinogens can cause cancer by causing mutations in the cellular
DNA. Such abnormal cells, if they are then stimulated to proliferate, can ultimately
progress to malignancy.
Knowledge of the development of breast lobules thus provides a coherent
explanation of the experimental results obtained in the rats treated with carcino-
gens. In rats not allowed to complete a pregnancy, most of the lobules would be type
2, in which most cancers are known to arise due to lack of full differentiation.
Treating these rats with a carcinogen therefore resulted in most developing breast
cancer. In contrast, rats allowed to bear a full-term litter of pups were resistant to
the carcinogen, since most of their breast lobules had developed into types 3 and 4.

In other words, the carcinogen would have caused just as many mutations, but any
abnormal cells that may have been generated were incapable of proliferation, and
therefore incapable of becoming cancerous.
The facts about lobular development in the breast also provide a clear expla-
nation of the epidemiological data. That is, the completion of a full-term pregnancy
provides some level of permanent protection against breast cancer, because it leaves
a woman with fewer vulnerable, undifferentiated cells which can give rise to can-
cer. The younger a woman is when she has her first full-term pregnancy, the greater
16
J. Russo and I. H. Russo, “Susceptibility of the Mammary Gland to Carcinogenesis,”
American Journal of Pathology 100.2 (August 1, 1980): 497–512.
THE NATIONAL CATHOLIC BIOETHICS QUARTERLY \ SUMMER 2005
308
Lobules Type 1
Lobules Type 3
BRIND \ THE ABORTION–BREAST CANCER CONNECTION
309
the protection, since it means there would be less time overall during which her
breasts contained a large percentage of such vulnerable cells. The breast cell situa-
tion with induced abortion is that not only are the cells not yet differentiated, but
because of the growth stimulation of pregnancy hormones—mainly estradiol—dur-
ing the incomplete pregnancy, there are more of those cancer-vulnerable cells in the
breasts than were there at the start of the pregnancy. Consequently, most epidemio-
logical studies have shown higher risks in women who have had an induced abortion
than in those who had not become pregnant at all.
Knowledge of the actions of estrogen and progesterone in terms of their ef-
fects upon breast growth completes the coherent picture of induced—but not spon-
taneous—abortion and breast cancer risk. As long as some progesterone is present
(called a “permissive hormone” in this situation), estradiol (which, recall, is a form of
estrogen) is a strong promoter of cellular proliferation in the breast. The greatest

growth stimulation occurs in the first and second trimesters of a normal pregnancy
by far. Importantly, estrogen is implicated in most known risk factors for breast
cancer; everything from taking artificial estrogens in the form of birth control pills or
postmenopausal estrogen replacement therapy, to beginning menstruation at an early
age and/or having menopause at a later age (both of which cause a greater lifetime
exposure to estrogen). Even non-reproductive risk factors such as postmenopausal
obesity and chronic alcohol consumption are explained in terms of such women
having higher chronic circulating levels of estrogen (because fat cells actually make
estrogen and alcohol impedes the liver’s ability to degrade estrogen).
Epidemiological Data Continue to Accrue
During the 1980s and early 1990s, as various researchers studied older and
older populations of women who had been exposed to legalized abortion, study after
study—in Japan,
17
Europe,
18
and the United States
19
—continued to report signifi-
cantly increased breast cancer risk in women who had had an induced abortion. By
1994, six epidemiological studies out of seven in the United States, on women of
both black and white ethnicity, had reported increased risk with induced abortion.
20
17
F. Nishiyama, “The Epidemiology of Breast Cancer in Tokushima Prefecture” [in
Japanese], Shikoku Ichi [Shikoku Medical Journal] 38 (1982): 333–343.
18
L. Lipworth et al., “Abortion and the Risk of Breast Cancer: A Case-Control Study in
Greece,” International Journal of Cancer 67.2 (April 10, 1995): 181–184.
19

H.L. Howe et al., “Early Abortion and Breast Cancer Risk among Women under Age
40,” International Journal of Epidemiology 18.2 (June 1989): 300–304; A. E. Laing et al.,
“Breast Cancer Risk Factors in African-American Women: The Howard University Tumor
Registry Experience,” Journal of the National Medical Association 85.12 (December1993):
931–939.
20
The six studies that reported a positive association were: M. C. Pike et al., “Oral
Contraceptive Use and Early Abortion as Risk Factors for Breast Cancer in Young Women,”
British Journal of Cancer 43.1 (January 1981): 72–76; L. A. Brinton et al., “Oral Contracep-
tives and Breast Cancer Risk among Younger Women,” Journal of the National Cancer
Institute 87.11 (June 7, 1995): 827–835; L. Rosenberg et al., “Breast Cancer in Relation to the
Occurrence and Time of Induced and Spontaneous Abortion,” American Journal of Epide-
THE NATIONAL CATHOLIC BIOETHICS QUARTERLY \ SUMMER 2005
310
It should be noted that most of these data reflected the standard of comparing the
effects of having a pregnancy which was ended by induced abortion versus the
effects of not having had that pregnancy (as opposed to versus continuing that
pregnancy to childbirth). Yet, even with the issue of breast cancer having had more
media exposure, and achieving major publicity with a “National Breast Cancer
Awareness Month” which highlighted that it had about 200,000 victims per year,
abortion—which was emerging as the most preventable of cancer causes—re-
ceived no attention whatsoever.
That changed abruptly at the end of October 1994, with the publication of a
study by Janet Daling and colleagues of the Fred Hutchinson Cancer Research
Center in Seattle, Washington.
21
The Daling study could hardly avoid a high profile,
as it was published in the Journal of the National Cancer Institute (JNCI). The
Daling team’s overall finding was of a statistically significant, 50 percent increase in
the risk of breast cancer among women who had chosen abortion. Even the New

York Times carried the story with the headline “New Study Links Abortions and
Increase in Breast Cancer Risk.”
22
But forces were already set in motion to make
sure the news was short-lived. For one thing, the Daling study was accompanied by
a most unusual JNCI editorial.
23
It was unusual because most medical journal edito-
rials, written by a scientist who has peer-reviewed the study, are published by the
journal in order to highlight the importance of a major study on a subject of wide
public interest. Such editorials typically make it easy for reporters—usually non-
scientists working on short deadlines—to glean the major points of a study and
render it understandable to the general public. Instead, Dr. Lynn Rosenberg, of
Boston University School of Medicine, took the opportunity to write an editorial
which sandbagged the Daling study, concluding—among other things—that “…the
overall results as well as the particulars are far from conclusive, and it is difficult to
see how they will be informative to the public.” Rosenberg did offer a possible
explanation for drawing the inferences that she did, introducing the idea that “re-
porting bias” could well have generated a false positive result. Since “reporting
miology 127.5 (May 1988): 981–989; H. L. Howe et al., “Early Abortion and Breast Cancer
Risk among Women under Age 40,” International Journal of Epidemiology 18.2 (June
1989): 300–304; A. E. Laing et al., “Breast Cancer Risk Factors in African-American Women:
The Howard University Tumor Registry Experience,” Journal of the National Medical Asso-
ciation 85.12 (December 1993): 931–939; A. E. Laing et al., “Reproductive and Lifestyle
Factors for Breast Cancer in African-American Women,” Genetic Epidemiology 11 (1994):
300. The one study that reported no association was: M. Moseson et al., “The Influence of
Medical Conditions Associated with Hormones on the Risk of Breast Cancer,” International
Journal of Epidemiology 22.6 (December 1993): 1000–1009.
21
J.R. Daling et al., “Risk of Breast Cancer among Young Women: Relationship to

Induced Abortion,” Journal of the National Cancer Institute 86.21 (November 2, 1994):
1584–1592.
22
Lawrence K. Altman, “New Study Links Abortions and Increase in Breast Cancer
Risk,” New York Times, Oct. 27, 1994, A24.
23
L. Rosenberg, “Induced Abortion and Breast Cancer: More Scientific Data Are
Needed,” Journal of the National Cancer Institute 86.21 (November 2, 1994): 1569–1570.
BRIND \ THE ABORTION–BREAST CANCER CONNECTION
311
bias” (also known as “response bias” or “recall bias”) continues to be employed as
the main reason to dismiss the ABC link, it will be discussed in some detail below.
But first, there are other aspects to the Daling study and its treatment in the
professional and popular media that need to be aired. In the study itself, it was
ominous enough that it showed that women in general suffered a 50 percent in-
creased risk of getting breast cancer after choosing abortion. However, there were
other findings yet more ominous. The risk was far more—more than a 100 percent
increase—for women who had an abortion prior to age eighteen or after age thirty.
The risk was also compounded for women who had any family history of breast
cancer—even a grandmother or aunt. That is, when women with a family history of
breast cancer and an abortion were compared with women with a family history of
breast cancer and no abortion, they were found to have an 80 percent increased
risk, rather than a 50 percent increased risk. As for women with the three risk
factors combined, that is, abortion before age eighteen with a positive family history
of breast cancer, the relative risk was actually reported as infinite. It should be
admitted that this last statistic was based on only twelve women; i.e., all twelve
women who had such a reproductive and family history were found among the 845
breast cancer patients and none of them were found among the 961 healthy control
women to whom they were compared. Could any other medical or surgical inter-
vention—especially one chosen by over a million healthy patients each year—that

raised such a specter of mortal danger in a major, peer-reviewed study, continue
unabated, and still be touted as safe?
Most assuredly not. But the number of abortion’s apologists and promoters,
within the ranks of science and medicine as well as politics and the media, is prodi-
gious. Rosenberg’s “poison pill editorial” in the JNCI gives us a hint, and ensuing
events were even more revealing. For example, the article in the New York Times,
while attracting attention by headlining Daling’s conclusion of overall increased risk,
was as dismissive as Rosenberg’s editorial. While it is appropriate to cut reporters
some slack for their being misled by the medical journal itself, it should be noted that
the author of the Times story, Dr. Lawrence Altman, was a seasoned epidemiologist
in his own right who worked for many years in the Centers for Disease Control. He
should certainly have known better. It turns out that Altman was later dragged
through the professional mud in July 1995, when an article in Science berated him
for giving the Daling study a credible headline.
24
Daling herself was treated even
worse, having her study belittled in the professional and lay media alike for several
months following its publication, although she continued to stand by her findings.
As for the editorialist Rosenberg, she could hardly claim any degree of objec-
tivity. Her editorial clearly misrepresented the Daling study, an act which by itself
satisfies most definitions of scientific misconduct. Rosenberg has also shown her-
self to take a stance that appears to go beyond “pro-choice” (as Janet Daling has
described herself), and that is radically pro-abortion. In 1999, for example, she served
on behalf of a group of Florida abortion clinics as an expert witness in their (ulti-
24
C. C. Mann, “Press Coverage: Leaving Out the Big Picture,” Science 269.5221 (July
14, 1995): 166.
THE NATIONAL CATHOLIC BIOETHICS QUARTERLY \ SUMMER 2005
312
mately successful) facial challenge of a new parental notification law in Florida for

minors seeking abortion.
25
Such minimal restrictions on abortion are supported by
the vast majority of even “pro-choice” citizens, but not by the likes of Rosenberg.
Other strange events bear her fingerprints. For example, a study on breast cancer in
South Australian women was published in the American Journal of Epidemiology
(AJE) in 1988,
26
when Rosenberg was an associate editor.
27
Only seven years
later, in the British Journal of Cancer, did a small review appear which revealed
that abortion—which had been omitted entirely as a variable in the 1988 AJE paper
(the data about abortion in that study was kept unpublished)—was the strongest and
most significant risk factor identified in the 1988 study.
28
Specifically, the 1988 data
showed that South Australian women who had had an induced abortion experienced
a statistically significant, 160 percent increased risk of breast cancer!
The Reporting Bias Canard
Of course, Rosenberg’s dismissive editorial that accompanied the Daling study
seemed plausible enough to the untrained reader, and she did offer a reason why
Daling’s results could not be trusted. What she wrote in this particular regard was:
“A major concern, especially because the observed effect was small, is the possibil-
ity of reporting bias.”
29
This attribution of an observed ABC link to reporting bias
(a.k.a. “response bias” or “recall bias”) refers to a potential weakness in any epide-
miological study that relies on a retrospective data collection method. What “retro-
spective” refers to is collecting data through interviews and/or questionnaires from

cancer patients and healthy control subjects who have been identified for the study.
Such studies rely, therefore, on the subjects’ ability and/or willingness to recall or
report their history of exposure to the variable in question. What the term “bias”
refers to in this context is a difference in the reporting accuracy between the cases
and the controls. If—the argument goes—breast cancer patients are more likely to
remember and report a history of abortion than are control subjects, then abortion
would appear more often among the patient population, even if no more of them had
had an abortion than the control women. This discrepancy would translate, in turn,
to an increased risk (i.e., a relative risk > 1.0) of breast cancer associated with
abortion, but it would not be a real result.
This reporting bias hypothesis seems all the more plausible because abortion is
such a sensitive subject for women to talk about. In fact, the underreporting of
25
North Florida Women’s Health and Counseling Services, Inc., et al., v. State of
Florida et al., Circuit Court, Second Judicial District, Leon County, FL, No. 99-3202, 1999.
26
T. E. Rohan, A. J. McMichael, and P. A. Baghurst, “A Population-Based Case-Control
Study of Diet and Breast Cancer in Australia,” American Journal of Epidemiology 128.3
(September 1988): 478–489.
27
Rosenberg was an associate editor of the journal during the period 1987 to 1991.
28
N. Andrieu et al., “Familial Risk, Abortion and Their Interactive Effect on the Risk of
Breast Cancer—A Combined Analysis of Six Case-Control Studies,” British Journal of Can-
cer 72.3 (September1995): 744–751.
29
L. Rosenberg, “Induced Abortion and Breast Cancer,” 1569.
BRIND \ THE ABORTION–BREAST CANCER CONNECTION
313
abortion history is well documented in the literature. The real question is whether or

not there is a difference in reporting accuracy between cases and controls. The
hypothesis of such a bias is certainly testable and worthy of testing. The problem
with the hypothesis, however, is that it has repeatedly been tested, and the bias
found not to exist.
Actually, the first paper that articulated the reporting bias argument in ABC
research, and claimed to have found such evidence, was a 1991 paper by a group
headed by Olav Meirik of the WHO.
30
However, a closer look at that claim reveals
just how far beyond the breaking point some are willing to stretch scientific credibil-
ity in order to reach a politically correct conclusion. The Meirik group had previously
performed a study (published in The Lancet)
31
on reproductive risk factors for
breast cancer in the late 1980s, based on retrospective interviews with women in
Sweden and Norway. The Swedish women in that study had also had their abortions
recorded at the time they took place (in addition to recalling them for the later study
interview), so in the 1991 AJE study, Meirik compared these prospective medical
records with the responses the women had given in the earlier study.
32
Meirik’s
group found a significant difference between the interview responses and the medi-
cal records, that is, “between underreporting of previous induced abortions among
controls [those who did not develop breast cancer] relative to overreporting among
cases [those who did develop breast cancer].”
33
This peculiar term, “overreporting,”
actually is intended to mean what it says; that is, that women who reported having
had an abortion of which the computer had no record were deemed to have imag-
ined the event! Without such a preposterous assumption, the Meirik group could

produce no significant evidence of reporting bias. Under some published peer pres-
sure—from both Daling’s group and my own—Meirik et al. quietly retracted the
claim of “overreporting” in a published letter in 1998.
34
However, they did not re-
tract the hypothesis of reporting bias in ABC research, and reporting bias continues
30
B. M. Lindefors-Harris et al., “Response Bias in a Case-Control Study: Analysis
Utilizing Comparative Data concerning Legal Abortions from Two Independent Swedish
Studies,” American Journal of Epidemiology 134.9 (November 1, 1991): 1003–1008.
31
Olav Meirik et al., “Oral Contraceptive Use and Breast Cancer in Young Women: A
Joint National Case-Control Study in Sweden and Norway,” The Lancet 2.8508 (September
20, 1986): 650–654; H O. Adami et al., “Absence of Association between Reproductive
Variables and the Risk of Breast Cancer in Young Women in Sweden and Norway,” British
Journal of Cancer 62.1 (July 1990): 122–126. Both papers report the results of the same
study, from somewhat different perspectives and using somewhat different statistical treat-
ments of the data.
32
B. M. Lindefors-Harris et al., “Response Bias in a Case-Control Study.”
33
Ibid.
34
Olav Meirik, Hans-Olav Adami, and Gunnar Eklund, letter to the editor, “Relation
between Induced Abortion and Breast Cancer,” Journal of Epidemiology and Community
Health 52.3 (March 1998): 209. See also Joel Brind et al., Reply to “Relation Between Induced
Abortion and Breast Cancer,” Journal of Epidemiology and Community Health 52.3 (March
1998): 209–211.
THE NATIONAL CATHOLIC BIOETHICS QUARTERLY \ SUMMER 2005
314

to be cited as fact by the NCI and all the other “mainstream” sources of public
health information.
Yet, not only is there no credible evidence for the existence of reporting bias in
ABC research, but there is ample credible evidence to prove its nonexistence by
any reasonable standard. Even going back to the early days of ABC research, a
1968 study published in Japan reported an increased risk of over 50 percent in
women who had had an abortion.
35
In this case, the fact that the control subjects
were also cancer patients—stomach cancer patients, rather than healthy controls—
is evidence against reporting bias. More recently, a 1989 study performed by the
New York State Department of Health, which was based on fetal death records—
not retrospective interviews—reported a 90 percent increase in risk among women
who had had an induced abortion (relative risk = 1.9).
36
In a more recent study in
2000, Daling’s group did something similar to what Meirik et al. had done.
37
They
compared retrospective interview responses with responses given by the same
women several years earlier, at the time of a prenatal interview, long before anyone
could have known who would end up with breast cancer. Their result was also
unequivocal: “The authors’ data do not suggest that controls are more reluctant to
report a history of induced abortion than are women with breast cancer.”
38
Though false, the reporting bias argument was—and still is—sufficient to keep
the ABC link as a credible scientific finding out of the public consciousness. As long
as the public keeps faith in government agencies such as the NCI, some voluntary
organizations such as the American Cancer Society, and the major medical journals
and societies, and so long as these agencies send a clear unified message, then it

does not matter whether the message is true or false: the public will believe it. The
only effective counterweight to such unified enforcement of the party line (in this
case, “safe abortion”) is an independent media. Importantly, the mainstream media’s
tendency to lean in the “pro-choice” direction is well known, and it reinforces their
tendency to view the mainstream scientific and medical authorities as authoritative.
The mainstream media is therefore not the prime mover in the ongoing effort to
deny the reality of the ABC link.
Comprehensive Review Refocuses Public Attention
This influence of the mainstream media became clear with the October 1996
publication of the “comprehensive review and meta-analysis” on the ABC link, which
I wrote in conjunction with Vernon Chinchilli, Walter Severs, and Joan Summy-Long
of the Pennsylvania State College of Medicine, and published in the British Medical
35
H. Watanabe and T. Hirayama, “Epidemiology and Clinical Aspects of Breast Can-
cer” [in Japanese], Nippon Rinsho 26.8 (August 1968): 1843–1849.
36
H. L. Howe et al., “Early Abortion and Breast Cancer Risk Among Women Under Age
40.”
37
Mei-Tzu C. Tang et al., “Case-Control Differences in the Reliability of Reporting a
History of Induced Abortion,” American Journal of Epidemiology 151.12 (June 15, 2000):
1139–1143.
38
Ibid.
BRIND \ THE ABORTION–BREAST CANCER CONNECTION
315
Association’s (BMA) epidemiology journal, the Journal of Epidemiology and
Community Health (JECH).
39
I had deliberately chosen this journal in order to get

a fair shake at publishing it without substantive change. Moreover, we did not wish
to have our study misrepresented and discounted in a JNCI editorial, as had hap-
pened to the 1994 Daling study. (As it happens, in December the JNCI did publish
an editorial which misrepresented and discounted our paper, but at least we had two
months for our message to gain some traction.
40
)
Our paper was both a narrative review and quantitative compilation of every
published study we could find (there were twenty-three) that had reported data on
the ABC link. We also enumerated the studies which reported data on spontaneous
abortion and those which did not distinguish between induced and spontaneous abor-
tion. Our conclusion was a statistically significant, overall 30 percent increase in the
risk of breast cancer among women who had had an induced abortion, and no
significant link with spontaneous abortion. Importantly, the comparison criterion for
induced abortion was—as it has always been—“not having had that pregnancy,”
rather than the childbirth choice. We also included an extensive discussion of the
theory and the evidence that form the biological basis of an ABC link.
Fortunately, the institutions involved in the publication of the meta-analysis
paper—Baruch College, Penn State Medical College, and the BMA—made efforts
to draw media attention, and the news was reported widely around the world. Un-
fortunately, but not surprisingly, the backlash from “mainstream” science and medi-
cine was swift and unrelenting.
Just three months after the publication of our meta-analysis, a new paper that
was widely hailed as the definitive disproof of the ABC link appeared in what is
arguably the most prominent and prestigious medical in the world, the New England
Journal of Medicine (NEJM).
41
Although this was a study only of women in
Denmark, and authored by Mads Melbye and colleagues of the Statens Serum
Institut in Copenhagen, it was largely funded by the U.S. Department of Defense,

and heralded by a NEJM editorial written by Patricia Hartge, a senior NCI scientist.
Hartge’s editorial conclusion was: “In short, a woman need not worry about the risk
of breast cancer when facing the difficult decision of whether to terminate a preg-
nancy.”
42
This followed quite naturally from the Melbye group’s conclusion: “In-
duced abortions have no overall effect on the risk of breast cancer.”
43
39
Joel Brind et al., “Induced Abortion as an Independent Risk Factor for Breast Cancer:
A Comprehensive Review and Meta-Analysis,” Journal of Epidemiology and Community
Health 50.5 (October 1996): 481–496.
40
D. L. Weed and B. S. Kramer, “Induced Abortion, Bias, and Breast Cancer: Why
Epidemiology Hasn’t Reached Its Limit,” Journal of the National Cancer Institute 88.23
(December 4, 1996): 1698–1700.
41
Mads Melbye et al., “Induced Abortion and the Risk of Breast Cancer,” New En-
gland Journal of Medicine 336.2 (January 9, 1997): 81–85.
42
Hartge, “Abortion, Breast Cancer, and Epidemiology.”
43
Melbye et al., “Induced Abortion and the Risk of Breast Cancer,” 81.
THE NATIONAL CATHOLIC BIOETHICS QUARTERLY \ SUMMER 2005
316
Just how, one may reasonably ask, could one single study’s result nullify al-
most half a century’s data from dozens of studies? The answer given was basically
a two-pronged argument: strength in numbers, and use of prospective medical records,
which eliminates the possibility of reporting bias. In regard to the former, the Melbye
study is no doubt the largest study on the ABC link, comprising every woman born in

the state of Denmark between 1935 and 1978 (over 1.5 million women), over 300,000
abortions and over 10,000 cases of breast cancer. The null result (relative risk =
1.00) was stated with very tight statistical limits (i.e., very high confidence). In the
public mind, the Melbye study pretty much did settle the issue, even though scientifi-
cally it was a complete travesty.
What was hailed as the Melbye study’s greatest strength actually turned out to
be its greatest weakness. That is, while its reliance on medical records does indeed
preclude the possibility of reporting bias, the flaws in the compilation of the data
were breathtaking. For starters, the study included records of induced abortions
dating back to 1973. Melbye’s justification for this seems straightforward enough:
“In 1973, the legal right to an induced abortion through 12 weeks’ gestation was
established for women with residence in Denmark.”
44
The implication here, of course,
is that Denmark had the equivalent of Roe v. Wade in 1973, before which time
elective abortion was not legal. In fact, induced abortion has been legal (and on
record) in Denmark for reasons other than medical necessity since 1939 and was
only most recently liberalized in 1973.
45
Consequently, some 60,000 of the oldest
women in the Melbye study cohort—the part of the population with the most cases
of breast cancer—who had had an abortion before 1973, were misclassified as not
having had an abortion, because their abortions had not been entered into the com-
puterized registry.
46
This is a very crucial point. It is not necessarily very conse-
quential if records are missing for many patients, providing these patients are thus
excluded from the analysis. But it is quite another matter when these patients are
included in the study as subjects, thus representing their abortion history as different
than it actually was.

Yet the Melbye study embodies still more egregious violations of proper scien-
tific methodology. Most glaring was the inclusion of breast cancer diagnoses dating
back to 1968. This actually violates the most fundamental rule of all in scientific
inquiry, i.e., temporality: cause must precede effect. Yet here, Melbye was measur-
ing the effect—breast cancer—from 1968, five years before beginning the collec-
tion of data in 1973 on the potential cause: induced abortion. Despite these glaring
gaps and flaws, the calculation of Melbye’s raw numbers resulted in a 44 percent
increase in breast cancer risk with induced abortion, an increase which did not
appear in print in the study, and which was made to disappear with statistical adjust-
44
Melbye et al., “Induced Abortion and the Risk of Breast Cancer,” 82.
45
M. Osler et al., “Family Planning Services Delivery: Danish Experience,” Danish
Medical Bulletin 37.1 (February1990): 95–105.
46
Befolkningens bevaegelser [Vital Statistics] 1994 (Copenhagen: Danmarks Statistik,
1996), 56-62.
BRIND \ THE ABORTION–BREAST CANCER CONNECTION
317
ment.
47
Even then, the Melbye study still reported—but not as a “conclusion”—a
statistically significant trend of increasing breast cancer risk with gestational age at
abortion. Thus, women who had aborted at more than eighteen weeks’ gestation
were at more than double the risk of breast cancer, compared to women who had
aborted at less than seven weeks gestation. Clearly, the Melbye study was purely a
political exercise in the guise of a scientific study, designed to shore up the reputa-
tion of abortion as safe for women.
While the Melbye study was generally quite effective in keeping the ABC link
out of public awareness, the issue kept rearing its head around the United States as

state after state considered abortion informed consent laws that mandated that ABC
link warnings be given by abortion practitioners to women seeking abortion. Many
such bills got as far as enactment into state law in the 1990s, specifically, those in
Mississippi, Montana (enjoined by a state court in 1999), Louisiana, and Kansas.
Texas and Minnesota followed in 2003. Over a dozen more states continue to de-
bate such legislation perennially. But the stubborn persistence of the issue has been
tested with a continuing program of denial from high places, most notably from the
National Cancer Institute.
Since the publication of the Daling study in 1994, the NCI has maintained a
“Cancer Facts” fact sheet devoted to abortion and breast cancer on its website.
Until early 1997, the message was generally dismissive of the ABC link, calling the
evidence for it “inconsistent and inconclusive.”
48
In 1997, however, shortly after the
publication of the Melbye study, the NCI ratcheted up the rhetoric of denial, claim-
ing that “there is no convincing evidence of a direct relationship between breast
cancer and either induced or spontaneous abortion.”
49
The inclusion of spontaneous
abortion here serves to confuse the issue, since the difference between induced and
spontaneous abortion in terms of pregnancy hormones and breast cancer risk had
long been resolved. But most outrageous about the new version of the NCI “fact
sheet” was the addition of the following sentence: “The scientific rationale for an
association between abortion and breast cancer is based on limited experimental
data in rats, and is not consistent with human data.” This of course, was exactly
contrary to reality, as the scientific rationale was based on many lines of evidence
(as discussed above), and the ABC link is entirely consistent with the overwhelming
majority of studies on women.
47
Joel Brind and Vernon M. Chinchilli, letter to the editor, “Induced Abortion and the

Risk of Breast Cancer,” New England Journal of Medicine 336.25 (June 19, 1997): 1834–1835.
48
National Cancer Institute, “Abortion, Miscarriage, and Breast Cancer Risk,” fact
sheet 3.75 (October 1996), formerly found at />Abortion_and_Breast_Cancer.html. This fact sheet is no longer available online, but a copy
may be obtained from Dr. Brind on request.
49
National Cancer Institute, “Abortion, Miscarriage, and Breast Cancer Risk,” fact
sheet 3.75 (February 1997), formerly found at />Abortion_and_Breast_Cancer.html. This fact sheet is no longer available online, but a copy
may be obtained from Dr. Brind on request.
THE NATIONAL CATHOLIC BIOETHICS QUARTERLY \ SUMMER 2005
318
Fortunately, there were some members of Congress who knew better, and
were also in a position to exercise their constitutional oversight authority to rein in
the NCI. Most notable was Dr. Tom Coburn (R-OK), an obstetrician as well as a
congressman, who managed to ask some hard questions of NCI representative Dr.
Edison Liu at a 1998 congressional subcommittee hearing in the House Commerce
Committee. Within months after a pointed follow-up letter was sent to the NCI
Director from the Chairman of the Commerce Committee, Tom Bliley (R-VA),
50
there was a substantial modification of the language of the NCI’s website “fact
sheet.” Thus, “no convincing evidence” morphed into “evidence of a direct relation-
ship is inconsistent.” Most importantly, the most offensive and untruthful state-
ment about the “scientific rationale” and the data’s “not [being] consistent with
human data” was completely expunged.
51
With the NCI forced to take at least a noncommittal stance, the ABC link
continued to emerge around the United States as a state issue in terms of informed
consent legislation.
A Slew of Large New Studies
Between the years 2000 and 2003, several new studies were published on

women from the United Kingdom, China, and Scandinavia, studies which, yet again,
were trumpeted as definitively disproving the ABC link. When subjected to close
scrutiny, however, the validity of their findings again runs into serious trouble.
The 2001 U.K. study was authored by Oxford epidemiologist Michael Goldacre
et al., and published in the JECH.
52
Interestingly—and it would seem hardly by
coincidence—the Goldacre study had overlapping authorship with both the 1982
Oxford study
53
and the 2004 “collaborative reanalysis.”
54
The Goldacre study was
widely considered to be very strong because, like the 1997 Melbye study, it involved
a very large number of women (over 350,000), over 28,000 of whom had developed
breast cancer. Also like the Melbye study, the Goldacre study relied entirely upon
medical records of abortion, from the (U.K.) National Health Service (NHS) hospi-
tal records. The results actually showed a statistically significant 17 percent de-
crease in breast cancer risk among women who had had any induced abortions.
However, also like the Melbye study, many missing abortion records resulted in the
misclassification of abortion-positive women as abortion-negative. But even more
egregiously than in the Melbye study, more than 90 percent of abortion-positive
women were misclassified in this way. This could be quite easily determined, since
50
Tom Bliley to Richard Klausner, January 12, 1999. A copy of this letter may be ob-
tained from Dr. Brind on request.
51
Joel L. Brind, “NCI’s New ABC ‘Facts’: Fewer Lies,” Abortion-Breast Cancer Quar-
terly Update 3.3 (Fall 1999): 1–2.
52

M. J. Goldacre et al., “Abortion and Breast Cancer: A Case-Control Record Linkage
Study,” Journal of Epidemiology and Community Health 55.5 (May 2001): 336–337.
53
Vessey et al., “Oral Contraceptive Use.”
54
Collaborative Group, “Breast Cancer and Abortion.”
BRIND \ THE ABORTION–BREAST CANCER CONNECTION
319
the Goldacre study was based on all the women who had been admitted to NHS
hospitals in the Oxford area for any reason. A simple perusal of statistics on induced
abortion in the United Kingdom reveals that at least 15 percent of those women
were abortion-positive,
55
yet the records upon which the Goldacre study relied indi-
cated that only just over 1 percent of the cancer patients—300 of them, to be ex-
act—had an induced abortion on record. Such a massive gap in the database ren-
ders any study of the ABC link in this population statistically meaningless.
56
China, on the other hand, comprising a population under very tight control, and
in which abortion is very common, would seem to offer a very good place to do
ABC link research. Between 2000 and 2002, three published reports of two studies
on overlapping populations of female textile workers in Shanghai appeared in the
American Journal of Epidemiology,
57
the International Journal of Cancer,
58
and the British Journal of Cancer.
59
These studies were quite large, reflecting a
study population of over 260,000 total, and the Ye study was based on prospective

data—that is, women had filled out questionnaires when they entered the study,
before any of them developed breast cancer. The Sanderson study was based on
retrospective interviews, and the results of both studies were essentially null, that is,
no significant effect on breast cancer risk was found associated with induced abor-
tion. The main problem with the Chinese studies is that abortion is so common (over
50 percent of the study population in both the Sanderson and Ye studies). In addition
(and contrary to the usual pattern in the United States and the United Kingdom),
almost all abortions are done after the birth of a woman’s first (and usually only)
child. Thus, the women in the study who had not had an abortion at a given age were
not typical. These women were either childless or they had their first child at an
older age, characteristics which are both risk factors for breast cancer. In other
words, the women who had an abortion were being compared to what amounts to a
high-risk subgroup of women, rather than to a group of women who are truly typical
of the population, and the risk-increasing effect of abortion was therefore masked.
60
55
Office for National Statistics (U.K.), “Report: Legal Abortions in England and Wales,
2000,” Health Statistics Quarterly 10 (Summer 2001): 57. As the overall induced abortion rate
in England and Wales averaged more than 1 percent per year during the study period (1968–
1998), it is conservatively estimated that approximately 15 percent of the women in the cohort
underwent an induced abortion in their lifetime.
56
Joel L. Brind and Vernon M. Chinchilli, letter to the editor, “Abortion and Breast
Cancer,” Journal of Epidemiology and Community Health 56.3 (March 2002): 237–238.
57
M. Sanderson et al., “Abortion History and Breast Cancer Risk: Results from the
Shanghai Breast Cancer Study,” abstract of presentation at the Society for Epidemiological
Research Annual Meeting, American Journal of Epidemiology 151.11 (June 1, 2000): 151.
58
M. Sanderson et al., “Abortion History and Breast Cancer Risk: Results from the

Shanghai Breast Cancer Study,” International Journal of Cancer 92.6 (June 15, 2001): 899–
905.
59
Z. Ye et al., “Breast Cancer in Relation to Induced Abortions in a Cohort of Chinese
Women,” British Journal of Cancer 87.9 (October 21, 2002): 977–981.
60
Joel L. Brind and Vernon M. Chinchilli, letter to the editor, “Breast Cancer and In-
duced Abortions in China,” British Journal of Cancer 90.11 (June 1, 2004): 2244–2246.
THE NATIONAL CATHOLIC BIOETHICS QUARTERLY \ SUMMER 2005
320
This difficulty is inherent in the nature of epidemiology, and it arises when the expo-
sure in question—in this case, induced abortion—has become the rule rather than
the exception within the population under study.
In Scandinavia, a study by Erlandsson et al. appeared in the International
Journal of Cancer in 2003.
61
This study was reasonably large—almost 1800 case-
control pairs, which made it about twice as large as the 1994 Daling study. As in
most of the other large recent studies discussed above, the data were collected
prospectively, and the study was widely viewed as more credible for that reason.
However, as in the Melbye and Goldacre studies, the Erlandsson study ran into the
misclassification problem resulting from huge gaps in the database. The subjects
were all Swedish women who had had at least one live birth during the study period,
because in Sweden, a record is automatically created at an antenatal interview. In
the antenatal record, each woman gives a detailed history, including any abortions.
The registry of antenatal record data was linked by Erlandsson et al. to the breast
cancer registry, in order to find any connection between induced abortion and breast
cancer. Similar to the Goldacre study, Erlandsson et al. found a 20 percent decrease
in risk of breast cancer with women who had had abortions, with a borderline sig-
nificance to that decrease. The problem here is that the typical pattern of induced

abortion in Sweden is more like that in China than the United States or the United
Kingdom; that is, abortion is used more often to limit family size than to delay first
childbirth. What that means for the Erlandsson study is quite simply that most of the
induced abortions in the study population happened sometime between the antenatal
interview (when all the abortion data were collected) and the time of breast cancer
diagnosis, and were therefore missing from the record. Here again, then, we find a
database which is simply unsuitable for obtaining a valid result regarding the ABC
link, because most of the women who had had an induced abortion were misclassified
as not having had one.
62
A final note regarding these large recent studies is in order. In the authors’
discussions of their findings in the text of the papers, all misrepresent the published
record of previous articles involving research on ABC. Specifically, all omit any
mention of the 1989 New York State study, a study which—prominently published in
the International Journal of Epidemiology—found an unequivocal ABC link us-
ing a rock-solid prospective database.
63
The most glaring example of this misrepre-
sentation appears in the Goldacre study: “None of the cohort or record linkage
studies have shown a significant increase in breast cancer risk after exposure to
induced abortion.”
64
61
G. Erlandsson et al., “Abortions and Breast Cancer: Record-Based Case-Control
Study,” International Journal of Cancer 103.5 (February 20, 2003): 676–679.
62
Joel L. Brind and Vernon M. Chinchilli, letter to the editor, “Abortions and Breast
Cancer: Record-Based Case-Control Study,” International Journal of Cancer 109.6 (May
10, 2004): 945–946.
63

Howe et al., “Early Abortion and Breast Cancer Risk among Women under Age 40.”
64
Goldacre et al., “Abortion and Breast Cancer: A Case-Control Record Linkage Study,”
337.
BRIND \ THE ABORTION–BREAST CANCER CONNECTION
321
More Changes on the NCI “Fact Sheet”
The rash of new, large studies which showed no ABC link set the stage for a
major change in the tone of the NCI “fact sheet” on the link. The change was set to
coincide with the arrival of new leadership in Washington, following the election of
Republican President George W. Bush. Just over a year into his first term, in early
March of 2002, his appointee for NCI Director, Andrew von Eschenbach, took the
agency’s helm. When he arrived at his new office, Dr. von Eschenbach found the
proposed text of the new ABC link “fact sheet” on his desk, waiting for his signa-
ture.
65
Two days later, the NCI’s message on the ABC link no longer referred to
evidence of the ABC link as “inconsistent.” Instead, the text read: “The current
body of scientific evidence suggests that women who have had either induced or
spontaneous abortions have the same risk as other women for developing breast
cancer.”
66
This presented a more conclusive tone than previous versions. The new
“fact sheet” was also shorter, the description of the 1994 study by Janet Daling and
colleagues having been eliminated entirely.
The net result was that the single most influential worldwide authority on what
causes cancer and what does not was firmly in denial of the ABC link. At the same
time, the underlings who orchestrate NCI policy had managed to tie what amounted
to an endorsement of “safe abortion” to the new conservative anti-abortion admin-
istration in Washington. The effect of the NCI policy change was chilling for those

engaged in the effort to raise public awareness of the ABC link. At the time the
change was made, there were two lawsuits in progress—one in North Dakota
67
and one in California
68
—which accused abortion providers of engaging in false
advertising by denying the ABC link. Armed with the new NCI “fact sheet,” judges
in both cases ruled against the plaintiff. In the North Dakota case, the judge actually
permitted the outrageous claim—made by the defendant—that there was “no evi-
dence” of the ABC link, to stand. In California, the judge dismissed the suit sum-
marily.
Politically, conservative voices in Washington were soon heard by the Bush
administration. Late in April, the National Physicians Center for Family Resources,
based in Birmingham, Alabama, sent Dr. von Eschenbach a detailed critique of the
NCI “fact sheet.” Among other flaws, the critique cited the NCI’s “overall tone of
65
A. von Eshenbach, opening address at NCI Workshop on Early Reproductive Events
and Breast Cancer, Bethesda, MD, February 24, 2003. Video record available at: http://
videocast.nih.gov/PastEvents.asp?c=1&s=91.
66
National Cancer Institute, “Abortion, Miscarriage, and Breast Cancer Risk,” fact
sheet 3.75 (March 6, 2002), formerly found at />This fact sheet is no longer available online, but a copy may be obtained from Dr. Brind on
request.
67
Amy Jo Mattson v. MKB Management Corp., dba Red River Women’s Clinic, District
Court, East Central Judicial District, County of Cass, State of North Dakota, Civil No. 99-3734.
68
Agnes Bernardo et al. v. Planned Parenthood Federation of America et al., Superior
Court of the State of California, County of San Diego, Case No. GIC 772552.
THE NATIONAL CATHOLIC BIOETHICS QUARTERLY \ SUMMER 2005

322
denial,” its “misrepresentation of the published medical literature on the ABC link,”
and its “inclusion of inaccurate statements.”
69
Then, in early June, a letter was sent to the Secretary of Health and Human
Services (HHS, the federal department which includes the NCI), Tommy Thomp-
son, by U.S. Congressmen Chris Smith (R-NJ), Joseph Pitts (R-PA), and twenty-
six other representatives, asking “that the fact sheet be reevaluated for accuracy
and bias and that it be removed from the Department website until that review is
completed.” They based this request on their stated belief that the “fact sheet” “is
scientifically inaccurate and misleading to the public.”
70
On July 1, 2002, the NCI removed the “fact sheet,” an action which prompted
predictable outrage from “safe abortion” advocates. A letter to HHS originating
from the office of Rep. Henry Waxman (D-CA) and signed by eleven of his fellow
“pro-choice” colleagues, referred to the ABC link as a “popular myth” and asked
for the restoration of the latest fact sheet.
71
Predictably, the “pro-choice” print
media rallied to Waxman’s cause, accusing the Bush administration of political med-
dling with “objective” science. One newspaper, the Star-Ledger (NJ), went so far
as to compare President Bush to Joseph Stalin as one out to control information and
hide the truth from the American people.
72
In late November, the removed fact sheet was replaced—sort of—with a
statement on “Early Reproductive Events and Breast Cancer.” The new statement
reverted to calling the data on the ABC link “inconsistent,” and announced an NCI
workshop to be conducted shortly, to help determine the current state of knowledge
and the direction of future research.
73

The NCI Workshop
In January 2003, the NCI announced its “workshop” on “Early Reproductive
Events and Breast Cancer” to be held in late February. It was billed as a three-day
workshop at which all the data on the ABC would be subject to “scientific scrutiny”
and “comprehensive review.” However, it was obvious from the start that the “work-
shop” was a scam. First, the meeting lasted barely two days—not three days, as it
began with Monday evening introductory addresses by Dr. von Eschenbach and
two other speakers, presentations of data on Tuesday morning, breakout sessions
69
National Physicians Center for Family Resources, “National Cancer Institute Fact
Sheet Analysis: The Abortion-Breast Cancer Connection (ABC Link),” fact sheet, April 2002.
A copy may be obtained from Dr. Brind on request.
70
Jocelyn Kaiser, “Cancer Risk: Nudge from Congress Prompts NCI Review,” Science
297.5579 (July 12, 2002): 171.
71
Office of Henry Waxman to Tommy Thompson, October 21, 2002, http://
www.house.gov/waxman/news_files/pdfs/news_letters_thompson_hhs_websites
_10_21_02.pdf.
72
Editorial, “Political Science,” The Star-Ledger (NJ), January 3, 2003, 16.
73
National Cancer Institute, “Early Reproductive Events and Breast Cancer,” http://
cancer.gov/cancer_information/doc.aspx?viewid=8cf78b34-fc6a-4fc7-9a63-6b16590af277.
BRIND \ THE ABORTION–BREAST CANCER CONNECTION
323
(off the record) on Tuesday afternoon, and a wrap-up and presentation of summary
findings by mid-day Wednesday.
74
Second, as if the time were not short enough, discussion of the ABC link was

severely limited, diluted by inclusion of animal research and basic reproductive hor-
mone and genetic research, as well as human reproductive research unrelated to
induced abortion. Even some of the invited experts who had published important
work on the ABC link, such as Janet Daling and the Russos, were invited to present
data only on breast cell and tumor genetics, and not on induced abortion. In fact,
most of the invited experts were not experts on the ABC link at all, and were
unaware of the fact that eradicating the ABC link was really the sole purpose of the
meeting. I myself was invited only at the eleventh hour, and only to have a seat in the
audience—not to make a presentation of any sort.
Third, it was particularly telling that the only experts who were invited to present
on the ABC link were publicly known to be on the side of discrediting it. There was
only one full-length, on-the-record lecture on the subject, given by Leslie Bernstein
of the University of Southern California,
75
and there were two off-the-record short
presentations of “late-breaking results,” one of them given by Mads Melbye. There
was neither any comprehensive review of extant published data, nor any opportu-
nity to review any of the new data, of which the audience was given only a “sneak
peek.”
76
I had no podium, but only the opportunity to ask a few questions. One
question I did ask was for access to Dr. Bernstein’s new data, and I was told that
she would not make it available until after it was published! At least this verbal
exchange was on the record.
77
Interestingly, one of Bernstein’s studies, which was
in press at the time of the “workshop,” was published a month later;
78
however, the
journal that published it took the unprecedented action of refusing to publish a letter

critical of the study from our group
79
—even though our group has standing in the
field of ABC link research.
If the NCI’s experts really wanted to settle the matter of the ABC link once
and for all, to debunk the extant published evidence and relegate it to the scrap heap
of epidemiological history, this was their chance to do so. The lone “crusader” (yours
74
National Cancer Institute, “Early Reproductive Events and Breast Cancer: Workshop
Agenda, February 24–26, 2003,” />75
Leslie Bernstein, address given at NCI Workshop on Early Reproductive Events and
Breast Cancer, Bethesda, MD, February 25, 2003. Video record available at: http://
videocast.nih.gov/PastEvents.asp?c=1&s=91.
76
Ibid.
77
Joel L. Brind, Q&A regarding Bernstein address, ibid.
78
Maya Mahue-Giangreco et al., “Abortion, Miscarriage, and Breast Cancer Risk of
Young Women,” Cancer Epidemiology Biomarkers & Prevention 12.3 (March 1, 2003):
209–214.
79
Joel Brind, and Vernon M. Chinchilli, letter to editor, “Re: Induced Abortion, Miscar-
riage, and Breast Cancer Risk of Young Women,” submitted to Cancer Epidemiology
Biomarkers & Prevention, July 10, 2003.
THE NATIONAL CATHOLIC BIOETHICS QUARTERLY \ SUMMER 2005
324
truly) was there, armed with all the data and opinion ever published in the peer-
reviewed literature on the subject. I could have been set up to present my analysis
of it to the assemblage, with the opportunity for all of them to take their best scien-

tific shots at it for as long as they liked. But they missed their chance. They played
out a political charade, just waiting to re-post a year later the conclusion that “the
best evidence” indicates “no overall association” between abortion and breast can-
cer on the NCI’s website. It was an exercise of brute political force, backed by the
power of the purse (our tax dollars), and any potential dissenters among the gath-
ered NCI grantees would assuredly have been held to account.
That brings up the fourth important point: the inherent conflict of interest be-
tween the NCI career scientists, who make the highly competitive grant funding
decisions for the agency that funds most of the cancer research in the United States
(and much of it elsewhere), and the supposedly independent research scientists who
receive those grants, that dare not break with the party line.
It is also noteworthy that political forces outside the NCI were keeping up
maximal pressure for the agency to arrive at the politically correct conclusion. A
month prior to the “workshop,” a heavy-handed New York Times editorial said it all:
“If the experts at the meeting agree that there is no link between abortion and breast
cancer, the institute will have no excuse to suppress the information. It will have to
issue a new fact sheet or admit it can no longer provide objective guidance on
matters that inflame social conservatives.”
80
As for von Eschenbach himself, his
willingness to play along with the NCI career scientists bent on discrediting the
ABC link may come as a surprise, because he was appointed by President Bush.
However, at the time of his appointment he was national president-elect of the
American Cancer Society (ACS), an organization with the same “safe abortion”
bias as the NCI. The ACS itself has engaged in a vigorous campaign against the
ABC link, including attempts to silence the free speech of the independent, non-
profit Coalition on Abortion/Breast Cancer, through legal intimidation.
81
Almost immediately after the “workshop,” the NCI posted its “summary find-
ings.”

82
The finding relevant to the ABC link reads: “Induced abortion is not associ-
ated with an increase in breast cancer risk (1)” (The “(1)” means that this conclu-
sion is accorded the highest quality of evidence rating, i.e., “well established”).
Interestingly, this “epidemiologic finding” was only the sixth out of ten, with the first
three relating to the “long-term risk reduction” in breast cancer risk attributable to
childbirth. The simplicity of connecting the proverbial dots to show that interrupting
a pregnancy in progress therefore leaves a woman at higher long-term breast can-
80
Editorial, “Abortion and Breast Cancer,” NewYork Times, Late Edition-Final, January
6, 2003, A20.
81
Coalition on Abortion/Breast Cancer, “Society’s Legal Eagles Harass the Coalition,”
/>82
National Cancer Institute, “Summary Report: Early Reproductive Events and Breast
Cancer Workshop,” />BRIND \ THE ABORTION–BREAST CANCER CONNECTION
325
cer risk, compared to not interrupting it, is exceeded only by the temerity of deliber-
ately disconnecting the dots.
As an invited participant to the NCI “workshop,” I felt compelled to file a
“minority report.”
83
One would think that, were the “workshop” not a sham, this
report would be posted on the NCI website. Instead, a brief excerpt is posted as a
“Minority Dissenting Comment,” without identifying the author or providing any
means to obtain the full text.
84
With the NCI giving its highest stamp of disapproval to the ABC link, convinc-
ing the public of its reality has been made difficult, to say the least. For example, two
informed consent laws enacted in 2003, one in Texas

85
and the other in Minne-
sota,
86
mandated that information on the ABC link be given to women considering
abortion. However, in 2004 the ABC link information was removed from the state
website by the governor of Minnesota under pressure from the state medical asso-
ciation.
87
Subsequently, the commissioner of health of Louisiana removed ABC link
information
88
that had been in that state’s mandated booklet since 1995.
89
It certainly did not help matters when, in March 2004, Beral’s “collaborative
reanalysis” appeared in The Lancet.
90
Not only did the Beral “reanalysis” arrive at
essentially the same conclusion as the NCI “workshop,” but it also claimed to be a
thorough review of the extant data. Another important similarity—revealed by close
83
Joel Brind, “Early Reproductive Events and Breast Cancer: A Minority Report,” March
10, 2003, />84
National Cancer Institute, “Minority Dissenting Comment Regarding Early Repro-
ductive Events and Breast Cancer Workshop,” />doc.aspx?viewid=15e3f2d5-5cdd-4697-a2ba-f3388d732642.
85
Woman’s Right to Know Act (September 1, 2003), Texas Health & Safety Code, Chap-
ter 171, subchapter A (Texas Legislative Council, 2004). Available under “Chapter 171. Abor-
tion” at />86
Women’s Right to Know Act (April 14, 2003), Minnesota Statutes 2004, Chapter 145,

sections 4241–4249 (Minnesota Office of the Revisor of Statutes, 2004). Available under
“145.4241” and subsequent section numbers at />145/.
87
Dr. Robert Meiches, chief executive officer, Minnesota Medical Association, to Min-
nesota Governor Tim Pawlenty and Diane Mandernach, Minnesota commissioner of health,
December 9, 2003. A copy of this letter may be obtained from Dr. Brind on request.
88
Louisiana Department of Health and Hospitals, Abortion: Making a Decision (Ba-
ton Rouge, LA: Louisiana Department of Health and Hospitals, n.d.), te.
la.us/offices/publications/pubs-1/Abortion-MakingaDecision04.pdf; Associated Press,
“Women Wrongly Warned Cancer, Abortion Tied,” November 9, 2004, />Health/wireStory?id=239655.
89
Louisiana Department of Health and Hospitals, Abortion: Making a Decision (Ba-
ton Rouge, LA: Louisiana Department of Health and Hospitals, 1995), 17. The 1995 version of
this document is no longer available from the Louisiana Department of Health and Hospitals,
but a copy may be obtained from Dr. Brind on request.
90
Collaborative Group, “Breast Cancer and Abortion.”
THE NATIONAL CATHOLIC BIOETHICS QUARTERLY \ SUMMER 2005
326
scrutiny—shows the “reanalysis” to be every bit as disingenuous as the “work-
shop.”
Although the group of “fifty-three epidemiological studies” (actually, fifty-two
with data specifically relating to induced abortion) comprising the “reanalysis” seems
impressive, it is not what it appears to be. With a total of forty-one studies extant in
the published literature, one naturally assumes that the “reanalysis” therefore in-
cluded data from eleven additional studies. However, this is far from the truth. What
Beral et al. actually did was exclude seventeen published studies, and add data from
twenty-eight that were previously unpublished. Two of the excluded studies were
excluded for appropriate scientific reasons, i.e., they had not ascertained the occur-

rence of abortion directly in cases and controls. But eleven of the studies were
excluded for entirely unscientific reasons, specifically, that “principal investigators
could not be traced,” or “original data could not be retrieved,” or “researchers
declined to take part in the collaboration,” or “principal investigators judged their
own information on induced abortion to be unreliable.” Four other studies were
excluded by simple omission, without any mention at all. The reader will hardly be
surprised that ten of these fifteen excluded studies had reported a statistically sig-
nificant ABC link. In fact, a compilation of all fifteen excluded studies reveals an
overall 80 percent risk increase among them. It is therefore hardly surprising that
the Beral group’s conclusion shows no significant risk increase, especially consider-
ing that twenty-eight of the fifty-two studies on which they relied had not been
previously published. Moreover, Beral et al. did include the large prospective studies
of the Melbye, Goldacre, and Erlandsson groups, studies which—as enumerated
above—should have been excluded on purely scientific grounds.
91
A Serious Methodological Flaw
Another important aspect of the Beral “reanalysis” is that it provides a par-
ticularly clear example of a standardized methodological flaw peculiar to ABC link
research, one which masks the most obvious connection between abortion and breast
cancer. In looking at the key data summary table in that study, it is important to note
the comparison that produced the overall negative result. The title of the table un-
equivocally states: “Relative risk of breast cancer, comparing the effects of having
a pregnancy that ended as an induced abortion versus effects of never having had
that pregnancy.” The artificiality of this comparison is striking, for “never having
had that pregnancy” is not an option for a woman already pregnant. In terms of any
reasonable standard of informed consent, the potential harm of any given medical or
surgical procedure must be weighed against the alternative of not having the proce-
dure. In the case of induced abortion—especially since this procedure is overwhelm-
ingly performed on healthy women—that would mean the childbirth alternative. As
far as breast cancer is concerned, the risk-reducing effect of full-term pregnancy

(FTP) has been well-known literally for centuries, and is universally acknowledged.
91
Breast Cancer Prevention Institute, “Fact Sheet: Abortion and Breast Cancer, Re:
‘Collaborative Reanalysis of Data’ Published in Lancet 3/25/05,” />beralpaperanalysis.htm.
BRIND \ THE ABORTION–BREAST CANCER CONNECTION
327
It is even acknowledged explicitly in the introduction to Beral’s “reanalysis”: “Preg-
nancies that result in a birth are known to reduce a woman’s long-term risk of breast
cancer.” It is hardly difficult to connect the dots here: Having an induced abortion
leaves a woman with a higher long-term risk of breast cancer, compared to not
having the abortion; i.e., compared to childbirth. Even ABC link nemesis Lynn
Rosenberg was forced to admit this under oath in cross-examination in court: “Ques-
tion: So in other words, a woman who finds herself pregnant at age fifteen will have
a higher breast cancer risk if she chooses to abort that pregnancy than if she carries
the pregnancy to term, correct?” Answer: “Probably, yes.”
92
It is therefore unequivocal that one aspect of the ABC link has long since been
proven beyond a reasonable doubt, but it has been hidden by a standard peculiarity
of epidemiological methodology. Yet even more telling is that this peculiarity reflects
a double standard, for it only seems to apply to abortion, while a more reasonable
standard is applied when evaluating other potential risk factors for breast cancer.
This double standard is exemplified most clearly when evaluation of the ABC link is
compared to evaluation of another risk factor in the news lately, i.e., postmeno-
pausal hormone replacement therapy (HRT). In fact, the parallel to HRT is really
quite striking, since FTP has effects similar to menopause on a woman’s long-term
breast cancer risk. In particular, both FTP and menopause lower a woman’s breast
cancer risk, and the younger the woman is when either event occurs, the more her
risk is lowered (i.e., the greater the protective effect).
Beral’s own collaborative group at Oxford actually helped to establish HRT as
a risk factor with their “Million Woman Study” on women in the United Kingdom,

published in The Lancet in 2003.
93
In conducting the study, the Beral group sent
questionnaires to women aged fifty through sixty-four. Importantly, not all women in
that age group have gone through menopause. It has also long been established that
women who go through menopause at a later age are at higher risk of breast cancer,
due to longer exposure to the cancer-promoting hormone estrogen. In fact, the
“million woman” study reproduced this very finding among all the non-users of
HRT: “The relative risk of breast cancer also varied substantially according to meno-
pausal status; for example, among never users of HRT the relative risk of invasive
breast cancer was 0.63 (0.58–0.68) for postmenopausal, compared with pre-
menopausal women.” (In other words, menopause reduced the risk of breast can-
cer by a statistically significant 37 percent, compared to premenopausal women of
the same age.) Therefore, including premenopausal women in the HRT analysis
would have only confounded the results. That is, comparing postmenopausal HRT
users to premenopausal women of the same age would make it look like HRT was
not a risk factor, because the risk among the premenopausal women is elevated in
the same manner by their bodies’ own estrogen production.
92
North Florida Women’s Health and Counseling Services, Inc., et al., v. State of
Florida et al., official transcript of videotape deposition of Lynn Rosenberg, Sc.D., for
purposes of trial testimony, Nov. 18, 1999, p. 77.
93
Million Woman Study Collaborators, “Breast Cancer and Hormone-Replacement
Therapy in the Million Woman Study,” The Lancet 362.9382 (August 9, 2003): 419–427.